Group F- Hackathon

Mission: Clean Growth

“Halve the energy use of new buildings by 2030”

Stakeholders: construction industry, home owners, private landlords, estate agents, housing associations, construction materials manufacturers, skills providers and national and local governments.

Research question

  1. What inequalities in building energy usage are there during periods of high and low temperature?

  2. What areas could be supported to achieve the potential energy performance ratings?

Research question context

During summer or winter, the more extreme temperatures cause people to use additional energy to either cool down or warm up their properties.

Low energy efficiency during such times would mean a lot of wasted energy to achieve that.

Identify regions more affected by the extreme temperatures that also have low energy efficiency/high energy consumption.

For such regions:

  1. What is the make-up of the properties?

  2. Are they mostly old dwellings, are they recently refurbished/insulated etc.?

  3. How much room is there to achieve the potential energy efficiency (does it correlate with dwelling age)?

Once these regions are identified and have enough potential for improvements:

  1. What could be the average costs of improving the energy efficiency of the dwellings within these regions ?

  2. whether they could reasonably be covered by homeowners and public funding needs to be provided?

Study area description

Focus on the City of London

Focus on June (summer month) and January (winter month)

Datasets

Energy usage statistics (gas, electric)

Building energy performance ratings

House building statistics

Fuel poverty

Building materials

OS INSPIRE Building Polygons

Dwelling Ages and Prices (LSOA)

Council Tax Band and Build Period by LA, MSOA, LSOA

House Prices for Small Areas (LSOAs)

Land Surface Temperature from MODIS (Moderate Resolution Imaging Spectroradiometer)

Spatial methods justification

Firstly, we can use clustering algorithms in the 3D area-energy efficiency space - are there concentrated zones of low/high efficiency? Can identify the optimal number of clusters from the cluster quality evaluation. Collapsing the clusters onto the 2D area and having the average energy efficiency determining the colour of the cluster would be a good way to answer this question.

Secondly, correlate the energy efficiency and temperature datasets both by rank and product-moment. Do zones of low energy efficiency also experience extreme temperatures?

Plotting spatial information by the LSOA

2017 annual average of Electricity consumption on LSOA scale.

2017 annual average of Gas consumption on LSOA scale.

Age of buildings (percentage of the buildings built post 2000)

Hottest temperatures during daytime in summer (June2018)

Coldest temperatures during daytime in winter (Jan2018)

Analysis and Results

We then did a KMeans clustering across 2 dimensions of space and the third dimension of energy consumption, and energy efficiency to find spatial locations with similar energy consumption patterns.

Histograms of energy consumption values along the three variables.

Sillouette Score of clustering obtained Optimal number of clusters = 4.

Spatial plot of the four clusters in space and within LSOA polygon spatial plots.

Histograms of energy effeciency values along the three variables.

Sillouette Score of clustering obtained Optimal number of clusters = 4.

Spatial plot of the four clusters in space and within LSOA polygon spatial plots.

Correlation analysis of indicators sorted by LSOA.

WINTER RESULTS

SUMMER RESULTS

Recommendations

1. The correlations between Night Temperatures in summer and winter with Energy consumption agreed with our expectation of increased energy consumption with extreme temperatures both in summer and in winter. That is, consumption increases with increasing temperatures in summer, and also increasing with decreasing temperatures in winter.

2. The temperatures didn’t correlate very well with the energy efficiency, which indicates potential to improve upon the net energy consumption by promoting a positive correlation between the hottest regions and the energy efficient building. This is a recommendation we can make to reduce energy usage.

3. Clustering analysis revealed that there exists quite clear division between Central London and everywhere else. A distinct cluster in the area-energy efficiency space was identified for both the current efficiency rating and the potential efficiency rating improvement. Only very few LSOAs of the higher efficiency cluster were located outside of Central London. We would recommend further investigation in how these particular LSOAs were developed and what could be applied to other areas of lower efficiency.

4. If all the suggested energy efficiency improvements are made for all the LSOAs, the total energy consumption savings would total 185,000 TWh.

Data recommendations

1. We found that there was enough data, but not enough time.

Team F thanks Bonnie, Obi, and Matt for organizing an amazing event!!